Aug 8 2023 STEM faculty session

Outline:

Topic	GAI elements	demos and examples
intro and overview	clarify goal is concrete toolkit expect students to be exposed, and some will be highly proficient re-assess course goals and learning objectives, at individudal, dept'l and concentration level. deconstruct each aspect of course adjust assessments, assignments, grade weighting need to incentivize level of engagement for deep learning. That means grades academic integrity aspects- don't set unenforcable policies. turn to Bok center for startup on GAI. scope-limited LLM queries plug-ins what's coming- notebook tool need to tell registrar if you want a final exam training data and settings address directly variation in familiarity. Need to retool. spectrum of use, see below faculty categories - resisting ignoring experimenting reflecting Embracing explicit expectations for LLM competencies
grounding principles & student expectations.	Students bear responsibility for all submitted work. 2. Cite all resources and collaborative work, incl GAI 3. Ethical and responsible use 4. respect privacy and confidentiality 5. Validation and Verification are essential	explicit examples would be useful.
Learning Goals	top level- adept and ethical users of these new tools Student capabilities and habits of mind topical material
Assessments	Two-part: open-resource take home, and in-class In-person final exam no calculator but sheet of notes. This implies midterm of same format Which in turn implies assignments with that element. in-class assessments that aren't tests
Assignments	adjust amount of GAI from all to none have no-calculator problems to be done without GAI tools, exams will draw upon this. real-time feedback on student understanding, change latency loop prospect of embedded-notebook with uploaded-materials access	friction code example ditch this one: 'act as a biology professor and produce four questions that will gauge my level of understanding of basic molecular biology" Then student answers them Then ask GAI to answer them Then critically assess 1) the questions, 2) your answers, 3) GAI answers (too sophisticated? too broad)
Lectures	shared-doc summaries and ranking free-narrative 'clicker' questions can it listen to lecture in real time? Have GAI ask questions? real-time determination of what was not clear have active-lecture group learning turn to GAI first! support structure.	demo doc summary
Use examples		show capture of chat history notebook examples turn off training
Syllabus	Need to establish and promulgate policy. See grounding principles	show OUE link and examples on Canvas
Course management and development	Syllabus generation Lecture outline generation Problems and assignments eventually- grading
Resources	Bok center prompt-generation courses https://learnprompting.org/ FAQ pages, OUE and divisional Case examples, curated somewhere.
(defer labs)

We should do a followon session on each of these topics, perhaps
accommodation aspect. OUE. scheduling curate a bank of examples.